Biochimica et Biophysica Acta, 389 (1975) 449-463
Elsevier Scientific Publishing Company, Amsterdam-Printed in The Netherlands BBA 76961
S I A L O G L Y C O P E P T I D E S F R O M BOVINE M I L K F A T G L O B U L E MEMBRANE ROGER HARRISON, JOHN D. HIGGINBOTHAM* and ROLAND NEWMAN** Biochemistry Group, School of Biological Sciences, University of Bath, Claverton Down, Bath (U.K.)
(Received December 13th, 1974)
SUMMARY Milk fat globule membrane was shown to contain sialic acid, all of which could be released without disruption of the fat globule. Sialoglycopeptides were cleaved from the surface of intact fat globules by Pronase and fractionated on Sephadex G-50. Further fractionation of the major sialoglycopeptide peak on DEAE-Sephadex gave two groups of sialoglycopeptides eluted with 0.1 M N a C I (Group A) and 0.5 M N a C I ( G r o u p B), respectively. Refractionation gave a major sialoglycopeptide from each of the two groups together with a total of three minor sialoglycopeptides. All five sialoglycopeptides eluted as single peaks using shallow salt gradients on DEAE-Sephadex and contained a hydrophilic peptide chain together with galactose, mannose, N-acetylgalactosamine, N-acetylglucosamine, and sialic acid. Glycopeptides of Group A but not G r o u p B contained fucose. The major sialoglycopeptide of Group B released 35 ~ of its hexose and hexosamine on treatment with alkaline borohydride leaving a sialoglycopeptide which had reduced serine and threonine and elevated alanine levels and in addition contained 2-aminobutyric acid. An oligosaccharide fraction containing N-acetylgalactosaminitol, galactose and sialic acid in a molar ratio of 1 : 1 : 2 was partially characterised from the cleavage mixture. The major sialoglycopeptide of Group A had a more complex carbohydrate structure and showed no released carbohydrate on treatment with alkaline borohydride. The sialoglycopeptides of milk fat globule membrane show many similarities with those of erythrocyte membrane and have a potential use in comparative and structural studies.
INTRODUCTION Available evidence indicates that during milk secretion the milk fat globule gains a membrane which is derived directly from the plasma membrane of the m a m m a r y * Present address: Tate and Lyle Ltd., Group Research and Development, Philip Lyle Memorial Research Laboratory, P. O. Box 68, Reading, RG6 2BX, U.K. ** Present address: Abt. Immunobiologie, Universit~it K61n, 5, K61n, 41, Kerpener Strasse 15, W. Germany.
450 cell [1-3] and possibly also from Golgi vesicle membrane [4, 5]. Recent electron-microscope studies [6, 7] suggest that despite gradual loss after leaving the secretory cell a considerable percentage of the original membrane remains on the surface of globules in freshly expressed cream. There is some evidence, therefore, that milk fat globule membrane can represent a convenient source of mammalian membrane material for structural studies. The protein content of milk fat globule membrane has been examined by several workers [1, 8, 9] and fractions containing hexose, hexosamine and sialic acid have been described [10, l 1]. However, little has been reported concerning the detailed composition and structure of purified glycoproteins. Recent microelectrophoretic studies [12] have shown that sialic acid, together with protein amine and carboxyl groups, contributes to the surface charge of milk fat globules, suggesting the presence of sialoglycoproteins on the membrane surface. We now report the isolation and partial characterisation of sialoglycopeptides enzymically cleaved from the surface of the milk fat globule membrane. A preliminary account of some of this work has been presented [13 ]. MATERIALS AND METHODS
Isolation of milk fat globules Uncooled milk from 12 Friesian cows in mid-lactation was pooled in an insulated container. Within 20 min of milking, the milk was warmed to 40 °C and separated in a bench cream separator (Alfa-Laval AE Farm Separator) calibrated to give 40 ",, w/v butter fat. The cream was diluted with three times its volume of double-distilled water at 40 0 C and reseparated. The dilution-separation process was repeated a further three times to give washed cream.
Preparation of milk fat globule membrane suspension Washed cream (100 ml) was cooled to 4 °C and shaken until butter formed. The buttermilk was separated from the butter, which was melted at 40 °C giving butte~ oil and butter serum. Combination of butter serum with buttermilk gave total milk fat globule membrane suspension (40 ml).
Analytical methods Evaporations were carried out in vacuo at 30-35 °C. Membrane filtration was performed using Oxoid Membrane filters (pore size 0.45/~m, Oxoid Ltd, Southwark Bridge Rd, London S.E.I.). Hexose and hexosamine. Total hexose was determined colorimetrically by a modification of the cysteine/HzSO4 assay [14]. For gas-liquid chromatographic analysis of individual hexoses and hexosamines in soluble carbohydrate-containing fractions, aliquots (0.1 ml) of solution (containing approx. 0.5 mg carbohydrate/ml) were diluted with an equal volume of 6 M HCI, saturated with nitrogen, sealed and heated at 100 °C for 4 h. The hydrolysates were cooled and an internal standard of mannitol (25/*g) was added to each sample. The solutions were then freed from excess H C I by azeotropic distillation with ethanol/ benzene (4 : 1, v/v). The residue in water (0.5 ml) was applied to a column (5 cm x 0.5 cm) of Bio-RadAG-50W ( × 8, H + form) resin. Neutral and basic components of
451
the hydrolysate were eluted with water (5 ml) and 2 M HCI (5 ml), respectively. The eluates were separately evaporated, freed from water and HCI by azeotropic distillation with ethanol/benzene (4 : 1, v/v) and converted into O-trimethylsilyl derivatives by incubation with Trisil (50 pl) (see Enzymes and chemicals ) at 37 °C for 15 min. Excess reagents were evaporated and the dry residue was reconstituted in n-hexane (25-200 #1). Aliquots (1-2 #1) were chromatographed on a column (6 ft x 0.25 inch) of coiled glass packed with 3 % silicone gum rubber SE-301, on High Performance (H.P.) Chromasorb W, 80-100 mesh. Samples were chromatographed isothermally at 175 °C with a nitrogen flow of 30.8 ml/min in a Perkin-Elmer F11 gas chromatograph equipped with a dual flame ionisation detector. Standard carbohydrate solutions were chromatographed immediately before and after each batch of samples to check detector response. Hexosamines were calculated as their N-acetyl derivatives. Fucose. Aliquots (0.1 ml) of carbohydrate-containing solution (approx. 0.2 mg carbohydrate/ml) were diluted with 0.025 M HzSO4 (0.4 ml) containing an internal standard of mannitol (I00 #g) and heated at 100 °C for 30 min. The hydrolysate was cooled to 0 °C and neutralised with saturated Ba (OH)z solution. BaSO4 was removed by centrifugation at 3000 x g for 3 min and the supernatant applied to a column (5 cm x 0.5 cm) of Bio-Rad AG-50W (X 8, H + form) resin. The aqueous eluate (5 ml) was evaporated, O-trimethylsilylated and analysed by gas-liquid chromatography as described for hexose and hexosamine but at 165 °C. 90 % recovery of fucose was obtained when a control sample was subjected to the above procedure. Sialic acid. For the determination of sialic acid in soluble carbohydrate-containing fractions, aliquots (0.1 ml) of solution (containing approx. 50 #g carbohydrate/ ml) were diluted with 0.1 M H/SO4 (0.1 ml) and heated in a sealed tube at 80 °C for 1 h. The cooled hydrolysate containing free sialic acid was applied to a column (5 cm x 0.5 cm) of Bio-Rad AG-I (X 8, formate form) anion-exchange resin as described by Svennerholm [15]. The resin was eluted successively with water (10 ml) and 0.3 M formic acid (10 ml). Formic acid was removed from the second eluate by azeotropic distillation with ethanol/benzene (4 : 1, v/v) and the resulting residue was finally reconstituted in water (0.2 ml) and assayed for sialic acid using the Aminoff [16] modification of Warren's [17] thiobarbituric acid method. 97 % recovery of N-acetylneuraminic acid was obtained when a control sample was subjected to the above procedure. Sialic acid was routinely calculated as N-acetylneuraminic acid which was the only sialic acid chromatographically detected after neuraminidase treatment of milk fat globules [12]. Sialic acid was released from the milk fat globule membrane either by dilute acid or neuraminidase treatment. In the former procedure washed cream (2 ml) was diluted with water (1 ml) and 0.2 M H2SO4 (5 ml) and heated at 80 °C for 100 min. These conditions gave maximal release of sialic acid without rupture of the milk fat globules. The cream was removed by centrifugation at 3000 x g for 30 min and the subnatant membrane-filtered before assay for free sialic acid as described above. Neuraminidase release of sialic acid was effected by incubation of washed cream (2 ml) with enzyme (1 mg) in 0.05 M barbiturate buffer, pH 5.5 [12] (10 ml) at 37 °C for 3 h. Cream was removed by centrifugation and the subnatant membrane-filtered and assayed for free sialic acid as above. Direct microscopic observation following both acid and neuraminidase treatment confirmed that the fat globules were intact. Total sialic acid content of disrupted membranes was determined by a combi-
452
nation of acid and extensive Pronase treatment of a membrane suspension. Milk fal globule membrane suspension (5 ml) was diluted with 0.2 M H2SO ~ ( 10 ml) and heated at 80 °C for 100 min. The cooled hydrolysate was centrifuged at 3000 x g for 15 rain and the supernatant dialysed (3 x 100ml water, 2 h each at 4 ~C). The diffusate was concentrated, membrane-filtered and assayed for free sialic acid as described above. The non-diffusible material was concentrated and incubated with 2 vol. of Ca 2 + buffer ( 4 r a M CaCI 2 in 0.05 M Tris-HCl buffer, pH 7.8) and 1 vol. of buffered pronase (0.05 M Tris-HC1 buffer, pH 7.8, containing 150/~g pronase/ml) at 37 ':C for 7 h. The cooled mixture was centrifuged at 3000 x g for 15 rain, membrane-filtered and assayed for sialic acid as described for soluble carbohydrate-containing fractions. The total sialic acid content of the membrane was taken as the sum of assays for the diffusate and non-diffusible material. Amino acids. Sialoglycopeptide solutions (50 HI) were saturated with nitrogen and hydrolysed in 6 M HCI in a sealed tube at 100 °C for 24 h. Samples were membrane-filtered and acid was removed by azeotropic distillation with ethanol/benzene (4 : l, v/v). An internal standard of DL-norvaline was added to the residue and analysis was carried out using a Technicon TSM auto analyser. Protein. Total protein content of fractions obtained from alkaline BH4 cleavage experiments was assayed by differential absorption at 215 and 225 nm as described by Waddell [18].
Enzymes and chemicals Neuraminidase (EC 3.2.1.18) from Clostridium perfringens, trypsin (EC 3.4.21.4) from bovine pancreas and N-acetylneuraminic acid were from Sigma London Ltd, Kingston-upon-Thames, Surrey. Pronase (B. grade, Streptomyces griseus protease) was from Calbiocbem. Ltd, 10, Wyndham Place, LondonW. I. Trisil (hexamethyldisilazane and trimethylchlorosilane in dry pyridine) is a product of the Pierce Chemical Co. Ltd and was purchased from Pierce and Warriner Ltd, Chester, Cheshire. Unless otherwise stated all other laboratory reagents were from BDH Ltd, Poole, Dorset. EXPERIMENTAL AND RESULTS
Siafic acid content of milk fat globule membrane The total sialic acid content of milk fat globule membranes was released by subjecting a suspension of membrane fragments to a combination of extensive Pronase and mild acid treatment (Materials and Methods) when 110~ 10/Jg sialic acid were obtained from the membranes corresponding to I ml washed cream. Similar amounts of sialic acid (I 15+4/~g/ml washed cream) were also obtained by treatment of milk fat globules with either mild acid or neuraminidase apparently without disruption of the globules. This suggests that the sialic acid is all present on the outer surface of the membrane although penetration of the membrane by neuraminidase or acid cannot be discounted.
E~ect of washing milk fat globule membrane Washed cream was gently stirred for 30 min with media of low and of high ionic strength (distilled water and 1.0 M Na C1) and also of varying pH (2-11). The result-
453 ing suspensions were centrifuged at 3000 x 9 for 3 min and the subnatants analysed for sialic acid-containing material. No significant quantities of sialic acid were released by any of the washing media. Extrinsic or absorbed proteins are normally removed by such washing techniques [19] and it appears therefore that the sialic acid of the milk fat globule membrane is not derived from adsorbed serum glycoproteins.
Proteolytie release of sialoglycopeptides from milk fat 9lobule membrane Trypsin. Washed cream was suspended in 4 vol. of buffered trypsin (0.01 M NaH2PO4 adjusted to pH 7.8 with 0.1 M NaOH, containing 50 ~lg trypsin/ml) and incubated at 37 °C. Release of sialoglycopeptides was followed by periodic removal of aliquots, centrifugation at 3000 x 9 for 3 min, membrane-filtration of the subnatant and assay for bound sialic acid (Materials and Methods). After 30 and 60 min incubation 23 ~ and 40 ~ , respectively, of the total membrane sialic acid had been released as sialoglycopeptides. Microscopic observation indicated extensive disruption of milk fat globules after 60 min incubation and this was confirmed by the appearance of butter oil. Pronase. Washed cream was suspended in 2 vol. of Ca 2+ buffer (4 mM CaC12 in 0.05 M Tris. HC1 buffer, pH 7.8) and incubated at 37 °C with 1 vol. of buffered pronase (0.05 M Tris. HC1 buffer, pH 7,8, containing 150/~g pronase/ml.) Release of sialoglycopeptides was determined as above for trypsin incubations. Approx. 40 ~ of the total membrane content of sialic acid was released after 30 rain without visible disruption of the milk fat globules which became evident only after 60 rain incubation. After 4 h incubation 100 ~ of the total sialic acid content of the milk fat globule membrane had been released and this was accompanied by total disruption of the fat globules and extensive butter oil formation.
Preparative-scale release of sialoglycopeptides from milk fat 91obule membrane Conditions were chosen so as to obtain the maximum yield of sialoglycopeptides with the minium of fat globule disruption and in view of the above experiments Pronase was preferred to trypsin for this purpose. Washed cream (500 ml), Ca 2 + buffer (1000 ml) and buffered pronase (500 ml) were separately warmed to 37 °C, mixed and incubated at 37 °C for 30 min with gentle stirring. Microscopic observation of the suspension after this treatment gave no evidence of milk fat globule disruption. The digest was cooled in ice and centrifuged at 3000 x 9 for 30 min to give a subnatant which was concentrated (500 ml) and extensively dialysed against distilled water at 4 °C. The non-diffusible material was concentrated (10 ml) and the resulting creamcoloured suspension centrifuged at 125 000 x # for 2 h. A pale yellow solution resulted which contained 18 mg hexose and 21 mg sialic acid. The solution was further concentrated to 2 ml.
Fractionation of Pronase-eleaved fragments on Sephadex G-50 The above concentrate (2 ml) was applied to a column (40 c m x 2.5 cm) of Sephadex G-50 (fine) (Pharmacia (OB) Ltd, Uxbridge Rd, London W.5), previously equilibrated with water at 4 °C. The column was eluted with water at 4 °C with a flow rate of 10 ml/h. Fractions (3.5 ml) were automatically collected and samples analysed for hexose and sialic acid. A 28o ,m was continuously monitored using an LKB Uvicord 2. The elution patterns
454 ~!
l
2
1,0~
0
,,
~.~
3
!,'1,
/,,'
100
ml
200
Fig. I. Gel filtration of pronase -cleaved glycopeptides on Sephadex G-50. The dialysed, centrifuged pronase digest (Experimental and Results) was concentrated and applied to a column (40 cm . 2.5 cm) of Sephadex G-50 equilibrated with water at 4 °C. Elution was with water at 4 "C with a flow rate o f 10 ml/h. Fractions (3.5 ml) were collected and analysed for hexose ( A 4 2 o . . . . - - - ) and sialic acid (A55o .m, ). A28o nm (.. -} was monitored using an LKB Uvicord 2.
are shown in Fig. 1. A major peak (Peak 2) contained approx. 90 ,'..~,of the sialic acid applied to the column and about 70 ~ of the hexose. Calibration of the column using proteins of known molecular weight (bovine serum albumin, a-chymotrypsin, myoglobin and insulin) gave a mean molecular weight for this partially included peak of approx. 8000 with a spread from 7000-9000. Only very low Az8 o .m was associated with this fraction, indicating a low content of aromatic amino acids. A minor peak (Peak 1), eluted in the void volume, contained the remaining sialic acid and some hexose, while a small hexose-containing peak (Peak 3), free from sialic acid, followed the major peak. Fractionation on DEAE-Sephadex A-25 o f Peak 2 glycopeptides Fractions corresponding to the major peak (Peak 2, Fig. 1) from Sephadex G-50 fractionation of Pronase-cleaved glycopeptide fragments from two batches (2 x 500 ml) of washed cream were combined, concentrated (150 ml), dialysed extensively against water at 4 °C and further concentrated (2 ml). The concentrate was applied to a column (40 cm x 2.5 cm) of DEAE-Sephadex A-25, previously equilibrated with 0.05 M Tris-HC1 buffer (pH 7.8) at 4 °C. The column was successively eluted at 4 ~C with buffer (85 ml), 0.1 M NaCI in buffer (85 ml) and 0.5 M NaCI in buffer (150 ml). Fractions (3 ml) were automatically collected and samples analysed for hexose and sialic acid. The elution patterns are shown in Fig. 2. Two major peaks, 2A and 2B, eluted with 0.1 M and 0.5 M NaCI, respectively, contained all the sialic acid and most of the hexose applied to the column. The hexose was distributed roughly equally between the two peaks while the sialic acid was divided in the ratio of approx.2 : 1 between Peaks 2B and 2A, respectively. The ratio varied slightly with different batches of cream. In addition to the two major fractions some sialic acid-free, hexose-containing material was eluted with buffer alone. Fractions corresponding to Peak 2A (Fig. 2) were collected, extensively dialysed against water at 4 °C, concentrated and refractionated on a column (40 c m x 1 cm) of DEAE-Sephadex A-25 as before except that a linear salt gradient (0-0.1M
455
L
,
2A
2B
,
2.0 c
. . . . . . . .
Elsevier Scientific Publishing Company, Amsterdam-Printed in The Netherlands BBA 76961
S I A L O G L Y C O P E P T I D E S F R O M BOVINE M I L K F A T G L O B U L E MEMBRANE ROGER HARRISON, JOHN D. HIGGINBOTHAM* and ROLAND NEWMAN** Biochemistry Group, School of Biological Sciences, University of Bath, Claverton Down, Bath (U.K.)
(Received December 13th, 1974)
SUMMARY Milk fat globule membrane was shown to contain sialic acid, all of which could be released without disruption of the fat globule. Sialoglycopeptides were cleaved from the surface of intact fat globules by Pronase and fractionated on Sephadex G-50. Further fractionation of the major sialoglycopeptide peak on DEAE-Sephadex gave two groups of sialoglycopeptides eluted with 0.1 M N a C I (Group A) and 0.5 M N a C I ( G r o u p B), respectively. Refractionation gave a major sialoglycopeptide from each of the two groups together with a total of three minor sialoglycopeptides. All five sialoglycopeptides eluted as single peaks using shallow salt gradients on DEAE-Sephadex and contained a hydrophilic peptide chain together with galactose, mannose, N-acetylgalactosamine, N-acetylglucosamine, and sialic acid. Glycopeptides of Group A but not G r o u p B contained fucose. The major sialoglycopeptide of Group B released 35 ~ of its hexose and hexosamine on treatment with alkaline borohydride leaving a sialoglycopeptide which had reduced serine and threonine and elevated alanine levels and in addition contained 2-aminobutyric acid. An oligosaccharide fraction containing N-acetylgalactosaminitol, galactose and sialic acid in a molar ratio of 1 : 1 : 2 was partially characterised from the cleavage mixture. The major sialoglycopeptide of Group A had a more complex carbohydrate structure and showed no released carbohydrate on treatment with alkaline borohydride. The sialoglycopeptides of milk fat globule membrane show many similarities with those of erythrocyte membrane and have a potential use in comparative and structural studies.
INTRODUCTION Available evidence indicates that during milk secretion the milk fat globule gains a membrane which is derived directly from the plasma membrane of the m a m m a r y * Present address: Tate and Lyle Ltd., Group Research and Development, Philip Lyle Memorial Research Laboratory, P. O. Box 68, Reading, RG6 2BX, U.K. ** Present address: Abt. Immunobiologie, Universit~it K61n, 5, K61n, 41, Kerpener Strasse 15, W. Germany.
450 cell [1-3] and possibly also from Golgi vesicle membrane [4, 5]. Recent electron-microscope studies [6, 7] suggest that despite gradual loss after leaving the secretory cell a considerable percentage of the original membrane remains on the surface of globules in freshly expressed cream. There is some evidence, therefore, that milk fat globule membrane can represent a convenient source of mammalian membrane material for structural studies. The protein content of milk fat globule membrane has been examined by several workers [1, 8, 9] and fractions containing hexose, hexosamine and sialic acid have been described [10, l 1]. However, little has been reported concerning the detailed composition and structure of purified glycoproteins. Recent microelectrophoretic studies [12] have shown that sialic acid, together with protein amine and carboxyl groups, contributes to the surface charge of milk fat globules, suggesting the presence of sialoglycoproteins on the membrane surface. We now report the isolation and partial characterisation of sialoglycopeptides enzymically cleaved from the surface of the milk fat globule membrane. A preliminary account of some of this work has been presented [13 ]. MATERIALS AND METHODS
Isolation of milk fat globules Uncooled milk from 12 Friesian cows in mid-lactation was pooled in an insulated container. Within 20 min of milking, the milk was warmed to 40 °C and separated in a bench cream separator (Alfa-Laval AE Farm Separator) calibrated to give 40 ",, w/v butter fat. The cream was diluted with three times its volume of double-distilled water at 40 0 C and reseparated. The dilution-separation process was repeated a further three times to give washed cream.
Preparation of milk fat globule membrane suspension Washed cream (100 ml) was cooled to 4 °C and shaken until butter formed. The buttermilk was separated from the butter, which was melted at 40 °C giving butte~ oil and butter serum. Combination of butter serum with buttermilk gave total milk fat globule membrane suspension (40 ml).
Analytical methods Evaporations were carried out in vacuo at 30-35 °C. Membrane filtration was performed using Oxoid Membrane filters (pore size 0.45/~m, Oxoid Ltd, Southwark Bridge Rd, London S.E.I.). Hexose and hexosamine. Total hexose was determined colorimetrically by a modification of the cysteine/HzSO4 assay [14]. For gas-liquid chromatographic analysis of individual hexoses and hexosamines in soluble carbohydrate-containing fractions, aliquots (0.1 ml) of solution (containing approx. 0.5 mg carbohydrate/ml) were diluted with an equal volume of 6 M HCI, saturated with nitrogen, sealed and heated at 100 °C for 4 h. The hydrolysates were cooled and an internal standard of mannitol (25/*g) was added to each sample. The solutions were then freed from excess H C I by azeotropic distillation with ethanol/ benzene (4 : 1, v/v). The residue in water (0.5 ml) was applied to a column (5 cm x 0.5 cm) of Bio-RadAG-50W ( × 8, H + form) resin. Neutral and basic components of
451
the hydrolysate were eluted with water (5 ml) and 2 M HCI (5 ml), respectively. The eluates were separately evaporated, freed from water and HCI by azeotropic distillation with ethanol/benzene (4 : 1, v/v) and converted into O-trimethylsilyl derivatives by incubation with Trisil (50 pl) (see Enzymes and chemicals ) at 37 °C for 15 min. Excess reagents were evaporated and the dry residue was reconstituted in n-hexane (25-200 #1). Aliquots (1-2 #1) were chromatographed on a column (6 ft x 0.25 inch) of coiled glass packed with 3 % silicone gum rubber SE-301, on High Performance (H.P.) Chromasorb W, 80-100 mesh. Samples were chromatographed isothermally at 175 °C with a nitrogen flow of 30.8 ml/min in a Perkin-Elmer F11 gas chromatograph equipped with a dual flame ionisation detector. Standard carbohydrate solutions were chromatographed immediately before and after each batch of samples to check detector response. Hexosamines were calculated as their N-acetyl derivatives. Fucose. Aliquots (0.1 ml) of carbohydrate-containing solution (approx. 0.2 mg carbohydrate/ml) were diluted with 0.025 M HzSO4 (0.4 ml) containing an internal standard of mannitol (I00 #g) and heated at 100 °C for 30 min. The hydrolysate was cooled to 0 °C and neutralised with saturated Ba (OH)z solution. BaSO4 was removed by centrifugation at 3000 x g for 3 min and the supernatant applied to a column (5 cm x 0.5 cm) of Bio-Rad AG-50W (X 8, H + form) resin. The aqueous eluate (5 ml) was evaporated, O-trimethylsilylated and analysed by gas-liquid chromatography as described for hexose and hexosamine but at 165 °C. 90 % recovery of fucose was obtained when a control sample was subjected to the above procedure. Sialic acid. For the determination of sialic acid in soluble carbohydrate-containing fractions, aliquots (0.1 ml) of solution (containing approx. 50 #g carbohydrate/ ml) were diluted with 0.1 M H/SO4 (0.1 ml) and heated in a sealed tube at 80 °C for 1 h. The cooled hydrolysate containing free sialic acid was applied to a column (5 cm x 0.5 cm) of Bio-Rad AG-I (X 8, formate form) anion-exchange resin as described by Svennerholm [15]. The resin was eluted successively with water (10 ml) and 0.3 M formic acid (10 ml). Formic acid was removed from the second eluate by azeotropic distillation with ethanol/benzene (4 : 1, v/v) and the resulting residue was finally reconstituted in water (0.2 ml) and assayed for sialic acid using the Aminoff [16] modification of Warren's [17] thiobarbituric acid method. 97 % recovery of N-acetylneuraminic acid was obtained when a control sample was subjected to the above procedure. Sialic acid was routinely calculated as N-acetylneuraminic acid which was the only sialic acid chromatographically detected after neuraminidase treatment of milk fat globules [12]. Sialic acid was released from the milk fat globule membrane either by dilute acid or neuraminidase treatment. In the former procedure washed cream (2 ml) was diluted with water (1 ml) and 0.2 M H2SO4 (5 ml) and heated at 80 °C for 100 min. These conditions gave maximal release of sialic acid without rupture of the milk fat globules. The cream was removed by centrifugation at 3000 x g for 30 min and the subnatant membrane-filtered before assay for free sialic acid as described above. Neuraminidase release of sialic acid was effected by incubation of washed cream (2 ml) with enzyme (1 mg) in 0.05 M barbiturate buffer, pH 5.5 [12] (10 ml) at 37 °C for 3 h. Cream was removed by centrifugation and the subnatant membrane-filtered and assayed for free sialic acid as above. Direct microscopic observation following both acid and neuraminidase treatment confirmed that the fat globules were intact. Total sialic acid content of disrupted membranes was determined by a combi-
452
nation of acid and extensive Pronase treatment of a membrane suspension. Milk fal globule membrane suspension (5 ml) was diluted with 0.2 M H2SO ~ ( 10 ml) and heated at 80 °C for 100 min. The cooled hydrolysate was centrifuged at 3000 x g for 15 rain and the supernatant dialysed (3 x 100ml water, 2 h each at 4 ~C). The diffusate was concentrated, membrane-filtered and assayed for free sialic acid as described above. The non-diffusible material was concentrated and incubated with 2 vol. of Ca 2 + buffer ( 4 r a M CaCI 2 in 0.05 M Tris-HCl buffer, pH 7.8) and 1 vol. of buffered pronase (0.05 M Tris-HC1 buffer, pH 7.8, containing 150/~g pronase/ml) at 37 ':C for 7 h. The cooled mixture was centrifuged at 3000 x g for 15 rain, membrane-filtered and assayed for sialic acid as described for soluble carbohydrate-containing fractions. The total sialic acid content of the membrane was taken as the sum of assays for the diffusate and non-diffusible material. Amino acids. Sialoglycopeptide solutions (50 HI) were saturated with nitrogen and hydrolysed in 6 M HCI in a sealed tube at 100 °C for 24 h. Samples were membrane-filtered and acid was removed by azeotropic distillation with ethanol/benzene (4 : l, v/v). An internal standard of DL-norvaline was added to the residue and analysis was carried out using a Technicon TSM auto analyser. Protein. Total protein content of fractions obtained from alkaline BH4 cleavage experiments was assayed by differential absorption at 215 and 225 nm as described by Waddell [18].
Enzymes and chemicals Neuraminidase (EC 3.2.1.18) from Clostridium perfringens, trypsin (EC 3.4.21.4) from bovine pancreas and N-acetylneuraminic acid were from Sigma London Ltd, Kingston-upon-Thames, Surrey. Pronase (B. grade, Streptomyces griseus protease) was from Calbiocbem. Ltd, 10, Wyndham Place, LondonW. I. Trisil (hexamethyldisilazane and trimethylchlorosilane in dry pyridine) is a product of the Pierce Chemical Co. Ltd and was purchased from Pierce and Warriner Ltd, Chester, Cheshire. Unless otherwise stated all other laboratory reagents were from BDH Ltd, Poole, Dorset. EXPERIMENTAL AND RESULTS
Siafic acid content of milk fat globule membrane The total sialic acid content of milk fat globule membranes was released by subjecting a suspension of membrane fragments to a combination of extensive Pronase and mild acid treatment (Materials and Methods) when 110~ 10/Jg sialic acid were obtained from the membranes corresponding to I ml washed cream. Similar amounts of sialic acid (I 15+4/~g/ml washed cream) were also obtained by treatment of milk fat globules with either mild acid or neuraminidase apparently without disruption of the globules. This suggests that the sialic acid is all present on the outer surface of the membrane although penetration of the membrane by neuraminidase or acid cannot be discounted.
E~ect of washing milk fat globule membrane Washed cream was gently stirred for 30 min with media of low and of high ionic strength (distilled water and 1.0 M Na C1) and also of varying pH (2-11). The result-
453 ing suspensions were centrifuged at 3000 x 9 for 3 min and the subnatants analysed for sialic acid-containing material. No significant quantities of sialic acid were released by any of the washing media. Extrinsic or absorbed proteins are normally removed by such washing techniques [19] and it appears therefore that the sialic acid of the milk fat globule membrane is not derived from adsorbed serum glycoproteins.
Proteolytie release of sialoglycopeptides from milk fat 9lobule membrane Trypsin. Washed cream was suspended in 4 vol. of buffered trypsin (0.01 M NaH2PO4 adjusted to pH 7.8 with 0.1 M NaOH, containing 50 ~lg trypsin/ml) and incubated at 37 °C. Release of sialoglycopeptides was followed by periodic removal of aliquots, centrifugation at 3000 x 9 for 3 min, membrane-filtration of the subnatant and assay for bound sialic acid (Materials and Methods). After 30 and 60 min incubation 23 ~ and 40 ~ , respectively, of the total membrane sialic acid had been released as sialoglycopeptides. Microscopic observation indicated extensive disruption of milk fat globules after 60 min incubation and this was confirmed by the appearance of butter oil. Pronase. Washed cream was suspended in 2 vol. of Ca 2+ buffer (4 mM CaC12 in 0.05 M Tris. HC1 buffer, pH 7.8) and incubated at 37 °C with 1 vol. of buffered pronase (0.05 M Tris. HC1 buffer, pH 7,8, containing 150/~g pronase/ml.) Release of sialoglycopeptides was determined as above for trypsin incubations. Approx. 40 ~ of the total membrane content of sialic acid was released after 30 rain without visible disruption of the milk fat globules which became evident only after 60 rain incubation. After 4 h incubation 100 ~ of the total sialic acid content of the milk fat globule membrane had been released and this was accompanied by total disruption of the fat globules and extensive butter oil formation.
Preparative-scale release of sialoglycopeptides from milk fat 91obule membrane Conditions were chosen so as to obtain the maximum yield of sialoglycopeptides with the minium of fat globule disruption and in view of the above experiments Pronase was preferred to trypsin for this purpose. Washed cream (500 ml), Ca 2 + buffer (1000 ml) and buffered pronase (500 ml) were separately warmed to 37 °C, mixed and incubated at 37 °C for 30 min with gentle stirring. Microscopic observation of the suspension after this treatment gave no evidence of milk fat globule disruption. The digest was cooled in ice and centrifuged at 3000 x 9 for 30 min to give a subnatant which was concentrated (500 ml) and extensively dialysed against distilled water at 4 °C. The non-diffusible material was concentrated (10 ml) and the resulting creamcoloured suspension centrifuged at 125 000 x # for 2 h. A pale yellow solution resulted which contained 18 mg hexose and 21 mg sialic acid. The solution was further concentrated to 2 ml.
Fractionation of Pronase-eleaved fragments on Sephadex G-50 The above concentrate (2 ml) was applied to a column (40 c m x 2.5 cm) of Sephadex G-50 (fine) (Pharmacia (OB) Ltd, Uxbridge Rd, London W.5), previously equilibrated with water at 4 °C. The column was eluted with water at 4 °C with a flow rate of 10 ml/h. Fractions (3.5 ml) were automatically collected and samples analysed for hexose and sialic acid. A 28o ,m was continuously monitored using an LKB Uvicord 2. The elution patterns
454 ~!
l
2
1,0~
0
,,
~.~
3
!,'1,
/,,'
100
ml
200
Fig. I. Gel filtration of pronase -cleaved glycopeptides on Sephadex G-50. The dialysed, centrifuged pronase digest (Experimental and Results) was concentrated and applied to a column (40 cm . 2.5 cm) of Sephadex G-50 equilibrated with water at 4 °C. Elution was with water at 4 "C with a flow rate o f 10 ml/h. Fractions (3.5 ml) were collected and analysed for hexose ( A 4 2 o . . . . - - - ) and sialic acid (A55o .m, ). A28o nm (.. -} was monitored using an LKB Uvicord 2.
are shown in Fig. 1. A major peak (Peak 2) contained approx. 90 ,'..~,of the sialic acid applied to the column and about 70 ~ of the hexose. Calibration of the column using proteins of known molecular weight (bovine serum albumin, a-chymotrypsin, myoglobin and insulin) gave a mean molecular weight for this partially included peak of approx. 8000 with a spread from 7000-9000. Only very low Az8 o .m was associated with this fraction, indicating a low content of aromatic amino acids. A minor peak (Peak 1), eluted in the void volume, contained the remaining sialic acid and some hexose, while a small hexose-containing peak (Peak 3), free from sialic acid, followed the major peak. Fractionation on DEAE-Sephadex A-25 o f Peak 2 glycopeptides Fractions corresponding to the major peak (Peak 2, Fig. 1) from Sephadex G-50 fractionation of Pronase-cleaved glycopeptide fragments from two batches (2 x 500 ml) of washed cream were combined, concentrated (150 ml), dialysed extensively against water at 4 °C and further concentrated (2 ml). The concentrate was applied to a column (40 cm x 2.5 cm) of DEAE-Sephadex A-25, previously equilibrated with 0.05 M Tris-HC1 buffer (pH 7.8) at 4 °C. The column was successively eluted at 4 ~C with buffer (85 ml), 0.1 M NaCI in buffer (85 ml) and 0.5 M NaCI in buffer (150 ml). Fractions (3 ml) were automatically collected and samples analysed for hexose and sialic acid. The elution patterns are shown in Fig. 2. Two major peaks, 2A and 2B, eluted with 0.1 M and 0.5 M NaCI, respectively, contained all the sialic acid and most of the hexose applied to the column. The hexose was distributed roughly equally between the two peaks while the sialic acid was divided in the ratio of approx.2 : 1 between Peaks 2B and 2A, respectively. The ratio varied slightly with different batches of cream. In addition to the two major fractions some sialic acid-free, hexose-containing material was eluted with buffer alone. Fractions corresponding to Peak 2A (Fig. 2) were collected, extensively dialysed against water at 4 °C, concentrated and refractionated on a column (40 c m x 1 cm) of DEAE-Sephadex A-25 as before except that a linear salt gradient (0-0.1M
455
L
,
2A
2B
,
2.0 c
. . . . . . . .
